State-of-the-art generalisation research in NLP: a taxonomy and review

D Hupkes, M Giulianelli, V Dankers, M Artetxe… - ar** for zero-shot cross-lingual hate speech detection
I Bigoulaeva, V Hangya, I Gurevych… - Language Resources and …, 2023 - Springer
The goal of hate speech detection is to filter negative online content aiming at certain groups
of people. Due to the easy accessibility and multilinguality of social media platforms, it is …

Sparse shield: Social network immunization vs. harmful speech

A Petrescu, CO Truică, ES Apostol… - Proceedings of the 30th …, 2021 - dl.acm.org
With the rise of social media users and the general shift of communication from traditional
media to online platforms, the spread of harmful content (eg, hate speech, misinformation …

Investigating cross-lingual training for offensive language detection

A Pelicon, R Shekhar, B Škrlj, M Purver… - PeerJ Computer …, 2021 - peerj.com
Platforms that feature user-generated content (social media, online forums, newspaper
comment sections etc.) have to detect and filter offensive speech within large, fast-changing …

CoRAL: a context-aware croatian abusive language dataset

R Shekhar, M Karan, M Purver - arxiv preprint arxiv:2211.06053, 2022 - arxiv.org
In light of unprecedented increases in the popularity of the internet and social media,
comment moderation has never been a more relevant task. Semi-automated comment …

ConOffense: Multi-modal multitask Contrastive learning for offensive content identification

D Shome, T Kar - 2021 IEEE International Conference on Big …, 2021 - ieeexplore.ieee.org
Hateful or offensive content has been increasingly common on social media platforms in
recent years, and the problem is now widespread. There is a pressing need for effective …

[PDF][PDF] Crowdsourcing of Parallel Corpora: the Case of Style Transfer for Detoxification.

D Dementieva, S Ustyantsev, D Dale, O Kozlova… - CSW@ VLDB, 2021 - ceur-ws.org
One of the ways to fighting toxicity online is to automatically rewrite toxic messages. This is a
sequenceto-sequence task, and the easiest way of solving it is to train an encoder-decoder …